Two-dimensional finite element modeling of slab detachment using Arbitrary Lagrangian-Eulerian method: The physical possibility of slab tear beneath the East Sea (Japan Sea)

When relatively more buoyant crustal parts (e.g., continental crust, seamount, and mid-ocean ridge) start to subduct, the extensional regime becomes predominant, which can lead to slab detachment. Globally, the detached part of slab can drive reorganization of plates configuration. As a result of slab detachment, Hot mantle material beneath the lower plane of subducting slab rises through the gap in the slab.

The origin of Baekdu Mountain in Korean Peninsula is enigmatic, because it is intraplate volcanoes far from plate margin (i.e., 1,500 km from the Japan Trench), which requires the hypothesis beyond arc volcanisms. The Big Mantle Wedge (BMW) hypothesis suggested dehydration from the stagnant Pacific slab above the 660km phase boundary. Contrasting to the BMW, hot mantle upwelling through detached slab (i.e., slab tear) is another hypothesis. Since it is difficult to find a deep mantle plume beneath Baekdu Mountain and slab tear is observed in the 660km discontinuity in seismic tomography, previous study (Tang et al., 2014) suggested the possibility of upwelling through the slab gap. However, the physical plausibility of detachment of slab beneath the East Sea has not been tested numerically.

In this study, 2D fluid dynamical modeling for slab detachment was conducted by the Finite Element Method (FEM) provided by COMSOL Multiphysics which can solve fully-coupled equations of Stokes, continuity and energy. For the rheology of our model, we used composite viscosity based of diffusion and dislocation creeps.

We adopted Arbitrary Lagrangian-Eulerian Method (ALE) to efficiently handle problems which has material interface with largely distortion and even breakoff. A remeshing algorithm was introduced to reconstruct the material interface at a certain time, since the ALE can only be used when the mesh and the grid are exactly same, deformation of the interface is simple. We benchmarked Stokes terminal velocity depending on density, viscosity and size of sinking circle based on the ALE implemented to COMSOL.

Then, we conducted a preliminary simulation of slab detachment (slab tear) to explain slab tear within subducted Pacific plate beneath the East Sea. With a wide parametric spaces of viscosity, potential temperature, subduction velocity/angle, we confirmed that mantle upwelling through slab gap at 660 km depth below the Baekdu Mountain can be reproduced. This implies that we need to carefully compare upwelling through slab gap and its contrasting hypothesis (i.e., the BMW).